328 bulletin: museum of comparative zoology. 



distinguished as a dark spherule closely apposed to the latter, as is shown 

 in the cell reproduced in Figure 95. Duriug this process the distal cen- 

 trosome retains its original position and structure unmodified, but the 

 tail filament grows until it is five or six times as long as the diameter of 

 the body of the spermatid. 



We now return to consider the fate of the remnants of the interzonal 

 filaments, and the remnants of the polar portion of the spindle now rep- 

 resented by an indefinite mass of archoplasm (Fig. 90) lying near the 

 centrosorae, which is as yet undivided. The changes taking place syn- 

 chronously in these two structures are not always absolutely the same; 

 either one may outstrip the other. The remnants of the interzonal fila- 

 ments no longer lie in the long axis of the cell, but at one side, nearer 

 or farther from the centrosomo as the migration of the latter has pro- 

 gressed to a greater or less degree. During the division of the centrosome 

 and the migration of its parts, these remnants rapidly dwindle in size, 

 until they are represented by only a small homogeneous globule or vesicle 

 (Fig. 92), and finally disappear altogether. As I have already stated, 

 there is no evidence that they are cast out of the cell, and I believe that 

 the substance of which they consist is broken down and absorbed by the 

 surrounding cytoplasm, though doubtless still existing in the form of 

 minute granules, which may reassert themselves in the formation of the 

 acrosome. It is also possible that a part of this material may join the 

 persistent archoplasmic structure derived from the polar portion of 

 the spindle, as is suggested by the increase in size and irregular behavior 

 of the latter. The polar remnants, at the time of the disintegration of 

 the remnants of the interzonal filaments consist of a single small, roughly 

 spherical mass (Fig. 90); but shortly afterward — by the time the inner 

 centrosome has reached the nuclear membrane in its migration — two 

 such masses are to be seen, both distinctly larger than the original one ; 

 of these one lies on either side of the inward (centripetal) extension of 

 the axial filament (Fig. 93). These two bodies, however, are not 

 sharply outlined from the surrounding cytoplasm, and are not to be re- 

 garded as distinct structures surrounded by the latter, but rather as 

 concentrations of archoplasm, which, on account of their chemical compo- 

 sition, appear dense and deeply stained. The two masses continue to 

 increase somewhat in size, but retain their relative positions on either 

 side of the axial filament throughout the later progress of the metamor- 

 phosis. At this early stage (shown in Figure 93) I have never been 

 able to find more than two such bodies, although in the mature sperma- 

 tozoon, as later described, three or even four may frequently be dis- 



